textread facing unexpected characters. Gives an error.

I'm trying to read the following file using textread with the following command [names, types, x, y, answer] = textread('data.txt', ... '%s %f %s %d %d', 'headerlines', 2)

cg00000292  0.826063401  ATP2A1  16  28890100
cg00002426  0.178659348  SLMAP  3  57743543
cg00003994  0.041451237  MEOX2  7  15725862
cg00005847  0.381918846  HOXD3  2  177029073
cg00006414  NA  ZNF425;ZNF398  7  148822837
cg00007981  0.037822681  PANX1  11  93862594

however it gives me the following error

Trouble reading floating point number from file (row 5, field 2) ==> NA  ZNF425;ZNF398  7
148822837\n

I understand that textread is expecting a float but reaches the NA and gives the error, now my question is how to tell matlab to replace the unexpected with NaN and proceed instead of breaking the execution.

Help is appreciated :)

Answer 1

[per isakson]

Try

    >> cac = cssm
    cac = 
        {6x1 cell}    [6x1 double]    {6x1 cell}    [6x1 int32]    [6x1 int32]

where

    function    cac = cssm()
        fid = fopen( 'cssm.txt' );
        cac = textscan( fid, '%s%f%s%d%d', 'Delimiter','\t', 'TreatAsEmpty','NA' );
        fclose( fid );
    end

and cssm.txt contains the sample data of the question

[Theo]

Thanks. It works :)